The overall goal of this project is to determine whether proton nuclear magnetic resonance (NMR) brain imaging can detect pathological changes in the brains of patients with Alzheimer's disease and patients with multi-infarct dementia, and whether NMR measures can distinguish these groups from each other and from age-matched control subjects. A related goal is to discover whether different regional abnormalities detected by NMR imaging correspond to specific positive findings in the neurological and neuropsychological examinations, CT scan, and ultimately neuropathological study. In order to accomplish these goals, 30 patients with Alzheimer's disease, 30 patients with multi-infarct dementia, and 30 nondemented control subjects will be examined during the next 3 years. The first 10 subjects in each group will be examined in all 3 years in order to plot NMR measures over time. The expectation is that patients with Alzheimer's disease will show changes in years 2 and 3, whereas patients with multi-infarct dementia and nondemented subjects will not. Three-demensional proton NMR studies with a variety of pulse sequences will provide images of cortical and subcortical structures, and will make it possible to quantify brain and cerebrospinal fluid volumes in relation to intracranial capacity. Quantitative determination of three NMR parameters, proton spin density (p), T1 relaxation time, and T2 relaxation time, will be performed in 14 homologous brain regions in the two cerebral hemispheres. Clinical data for comparison with NMR data will consist of results from comprehensive neurological examinations and neuropsychological tests of memory, attention, language, praxis, and visual perception. The contribution of this work could be to establish NMR imaging as a method of noninvasive brain biopsy, having extensive clinical and research application.